Rubric Spans 是标签表示：用于短答案评分的联合 LLM 编码
Rubric Spans are Label Representations: Joint LLM Encoding for Short Answer Scoring
浏览论文内容
中文总结 AI 辅助
提出 RUSPAN 框架，将评分标准视为标签表示并联合编码问题、回答与标准级别，通过评分标准无关掩码实现零样本迁移，在六种语言的 ASAS 基准上超越基线，并在 PT-ASAG 上取得显著提升。
中文摘要 AI 辅助
自动短答案评分（ASAS）要求模型能够根据特定问题的标准对学生回答进行评分，同时保持高效性并能在不同评分标准集之间迁移。我们提出了 RUSPAN，一个以评分标准为条件的 ASAS 框架，将评分标准描述视为语义标签表示。RUSPAN 将问题上下文、学生回答和所有候选评分标准级别序列化为单一序列，然后从单次语言模型（LM）前向传播中产生的评分标准片段和整个序列的表示中，对级别进行列表式评分。我们进一步引入了 RUSPAN-RIM，其中一种评分标准无关掩码（Rubric-Independent Mask）防止评分标准片段相互关注，使得评分标准表示仅依赖于回答和问题上下文，并在训练过程中防止对评分标准模式的过拟合，以实现零样本迁移。在涵盖英语、德语和葡萄牙语的六个 ASAS 基准上，RUSPAN 在单基准评分上优于判别式和生成式基线，而带有位置重索引的 RIM 在 PT-ASAG（具有最强语言和评分标准结构组合偏移的保留基准）上带来了一致且显著的提升。
英文摘要
Automatic Short Answer Scoring (ASAS) requires models that can score student responses against question-specific criteria while remaining efficient and transferable across rubric sets. We propose RUSPAN, a rubric-conditioned ASAS framework that treats rubric descriptions as semantic label representations. RUSPAN serialises the question context, student answer, and all candidate rubric levels into a single sequence, then scores the levels listwise from the rubric-span and whole-sequence representations produced in a single LM pass. We further introduce RUSPAN-RIM, in which a Rubric-Independent Mask prevents rubric spans from attending to one another, making rubric representations depend only on the answer and question context and preventing overfitting to rubric patterns during training for zero-shot transfer. On six ASAS benchmarks spanning English, German, and Portuguese, RUSPAN improves mono-benchmark scoring over discriminative and generative baselines, while RIM with position reindexing delivers consistent and substantial gains on PT-ASAG, the held-out benchmark with the strongest combined language and rubric-structure shift.
发表机构
- DIPF | Leibniz Institute for Research and Information in Education（DIPF | 莱布尼茨教育与信息研究所）
- Centre for International Student Assessment (ZIB)（国际学生评估中心（ZIB））
- Goethe University Frankfurt（法兰克福歌德大学）
机构由 AI 辅助整理，请以论文原文为准。
